FDTD Based Dictionary Matrix for Sparsity-Based Through-Wall Radar Imaging

Abstract

Compressive sensing for through-wall radar imaging (TWRI) is a promising method to obtain a high-resolution image with limited number of measurements. The capability of the existing method in the framework of CS is limited due to the model error stemmed from the approximated signal model which does not consider multipath returns or only consider first-order interior wall multipath returns. In order to exploit various multipath returns, finite-difference time domain (FDTD) technique is used to obtain the scattered signal for each assumed target position and then to construct the exact forward scattering model. Then, sparse reconstruction is used to solve this linear inverse problem. Numerical results demonstrate that the proposed approach performs better at ghost suppression in the same condition of the signal-to-noise ratio (SNR).